This is the sixth blog in the series “Design Your First Salesforce AI Agent with Agentforce : From Sales Persona to Automation“.
As AI agent development becomes more accessible, it’s time to rethink who gets to build. With tools like Prompt Builder and Flow, cross-functional teams — not just developers — can now shape how agents behave. In this post, we’ll outline how roles across the business contribute to each phase of the agent lifecycle, and close with a practical framework to help you scale from your first agent to many.
🤝 Who Can Build: Rethinking Roles in AI Agent Design
One of the most exciting things about building agents with Agentforce is how accessible the process has become.
Thanks to no-code tools like Prompt Builder, Agent Builder, and Salesforce Flows, business analysts, operations leaders, and Salesforce admins can now play an active role in shaping how AI agents behave — from mapping user intents to writing natural language instructions.
Developers remain essential, especially for building complex integrations and reusable custom actions. But with the right collaboration, the broader team can now contribute meaningfully to the agent-building process, ensuring that what’s built aligns closely with real business needs.
What matters most is having a shared understanding of:
- The persona you’re building for
- The jobs to be done in their workflow
- The data and systems the agent needs to access
- The behaviors and guardrails that define how it should respond
In fact, much of the agent-building lifecycle — from defining topics and intents to testing behaviors — can be led by cross-functional teams working alongside IT.
These responsibilities are intended as a sample reference. Depending on your organization’s structure, roles, and Salesforce implementation, ownership may vary. In some teams, a Salesforce Admin may also act as the analyst; in others, business operations leaders may lead much of the planning. Use this table to guide collaboration — not as a rigid rulebook.
👉 For the purposes of this blog, this breakdown highlights the kind of cross-functional teamwork that makes agent development successful.
🏁 Conclusion: Build Agents That Drive Real Work
This methodology offers a practical, structured way to go from high-level needs to a working AI agent — grounded in real workflows and designed to support specific tasks.
By starting with the user persona and mapping out jobs to be done, you’ve created a foundation that connects business impact to automation logic. Through clearly defined topics, tested actions, scoped instructions, and real-world validation, you’ve built an agent ready to support daily operations.
From here, continue with the broader roadmap of use cases you identified earlier. Add new topics, expand capabilities, and iterate based on real usage. Each topic builds on the same foundation — enabling a modular, scalable approach to intelligent automation across your organization.
To build agents using this framework refer to this template –
Checkout the next blog in the series here.
Written by
